diff mbox series

[net,v2,2/2] net: phylink: Add module_exit_stub()

Message ID 20231221085109.2830794-3-yi.fang.gan@intel.com
State New
Headers
Series Fix phylink unloadable issue |

Commit Message

Gan, Yi Fang Dec. 21, 2023, 8:51 a.m. UTC 
  In delete_module(), if mod->init callback is defined but mod->exit
callback is not defined, it will assume the module cannot be removed
and return EBUSY. The module_exit() is missing from current phylink
module drive causing failure while unloading it.
Add module_exit_stub() in phylink for the module to be unloadable.

Fixes: eca68a3c7d05 ("net: phylink: pass supported host PHY interface modes to phylib for SFP's PHYs")
Cc: <stable@vger.kernel.org> # 6.1+
Signed-off-by: Gan, Yi Fang <yi.fang.gan@intel.com>
---
 drivers/net/phy/phylink.c | 1 +
 1 file changed, 1 insertion(+)

Comments

Andrew Lunn Dec. 21, 2023, 9:22 a.m. UTC | #1 
On Thu, Dec 21, 2023 at 04:51:09PM +0800, Gan, Yi Fang wrote:
> In delete_module(), if mod->init callback is defined but mod->exit
> callback is not defined, it will assume the module cannot be removed
> and return EBUSY. The module_exit() is missing from current phylink
> module drive causing failure while unloading it.

You are missing justification it is actually safe to remove
phylink. Maybe Russell King deliberately did not implement
module_exit() because it can explode in interesting ways if it was?

	Andrew
Gan, Yi Fang Jan. 4, 2024, 9:45 a.m. UTC | #2 
Hi Andrew,

The function phylink_init() is introduced by others.
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.6.y&id=eca68a3c7d05b38b4e728cead0c49718f2bc1d5a 
The module_exit() is added by referring to
https://elixir.bootlin.com/linux/latest/source/kernel/module/main.c#L738.
Since the macro function is rejected, I will send a V3.
Let's see if Rusell King has any comment. Thanks.


Best Regards,
Gan Yi Fang

> -----Original Message-----
> From: Andrew Lunn <andrew@lunn.ch>
> Sent: Thursday, December 21, 2023 5:23 PM
> To: Gan, Yi Fang <yi.fang.gan@intel.com>
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>; Andrew Halaney
> <ahalaney@redhat.com>; Javier Martinez Canillas <javierm@redhat.com>; John
> Stultz <jstultz@google.com>; Rafael J . Wysocki <rafael@kernel.org>; Jens Axboe
> <axboe@kernel.dk>; Russell King <linux@armlinux.org.uk>; Heiner Kallweit
> <hkallweit1@gmail.com>; David S . Miller <davem@davemloft.net>; Eric
> Dumazet <edumazet@google.com>; Jakub Kicinski <kuba@kernel.org>; Paolo
> Abeni <pabeni@redhat.com>; Marek Behún <kabel@kernel.org>;
> netdev@vger.kernel.org; linux-stm32@st-md-mailman.stormreply.com; linux-
> arm-kernel@lists.infradead.org; linux-kernel@vger.kernel.org; Looi, Hong Aun
> <hong.aun.looi@intel.com>; Voon, Weifeng <weifeng.voon@intel.com>; Song,
> Yoong Siang <yoong.siang.song@intel.com>; Lai, Peter Jun Ann
> <peter.jun.ann.lai@intel.com>; Choong, Yong Liang
> <yong.liang.choong@intel.com>
> Subject: Re: [PATCH net v2 2/2] net: phylink: Add module_exit_stub()
> 
> On Thu, Dec 21, 2023 at 04:51:09PM +0800, Gan, Yi Fang wrote:
> > In delete_module(), if mod->init callback is defined but mod->exit
> > callback is not defined, it will assume the module cannot be removed
> > and return EBUSY. The module_exit() is missing from current phylink
> > module drive causing failure while unloading it.
> 
> You are missing justification it is actually safe to remove phylink. Maybe Russell
> King deliberately did not implement
> module_exit() because it can explode in interesting ways if it was?
> 
> 	Andrew
Andrew Lunn Jan. 4, 2024, 1 p.m. UTC | #3 
On Thu, Jan 04, 2024 at 09:45:47AM +0000, Gan, Yi Fang wrote:
> Hi Andrew,
> 
> The function phylink_init() is introduced by others.
> https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.6.y&id=eca68a3c7d05b38b4e728cead0c49718f2bc1d5a 
> The module_exit() is added by referring to
> https://elixir.bootlin.com/linux/latest/source/kernel/module/main.c#L738.
> Since the macro function is rejected, I will send a V3.
> Let's see if Rusell King has any comment. Thanks.

Please don't top post when using linux kernel mailing lists.

> > You are missing justification it is actually safe to remove phylink. Maybe Russell
> > King deliberately did not implement
> > module_exit() because it can explode in interesting ways if it was?

You still need to explain why it is safe to implement it.

    Andrew
diff mbox series

Patch

diff --git a/drivers/net/phy/phylink.c b/drivers/net/phy/phylink.c
index 25c19496a336..823c9b43cd92 100644
--- a/drivers/net/phy/phylink.c
+++ b/drivers/net/phy/phylink.c
@@ -3725,6 +3725,7 @@  static int __init phylink_init(void)
 }
 
 module_init(phylink_init);
+module_exit_stub(phylink);
 
 MODULE_LICENSE("GPL v2");
 MODULE_DESCRIPTION("phylink models the MAC to optional PHY connection");